THE NEXT THING, IN PERSPECTIVEIndependent AI news

The latest AI news, in focus.

Official announcements and industry reporting. Follow the timeline of AI.

Today's news

News feed
AI NEWS × FOXAIFind your next step

From the latest news to tools that work for you.

Explore AI tools

Discover AI tools

Swipe to explore →
JAXModels & infrastructure

JAX is a Python numerical computing library designed for accelerators, specifically tailored for high-performance array computations and large-scale machine learning. It efficiently executes the same code across multiple hardware backends, including CPUs, GPUs, and TPUs.

AI DEVELOPMENT PLATFORM分类推荐榜首Pricing unverifiedView details
Open DesignDesign & 3D

OpenDesign is an open-source design workspace that can connect to local programming agents, use built-in model keys, or leverage cloud-based models, enabling the creation and editing of prototypes, web pages, presentations, and graphics within a design system; local software is free, while model services incur fees.

AI design tools分类推荐榜首Free entry availableView details
EvoMapAgents & automation

EvoMap is an infrastructure designed for the reuse of AI agent experiences. It organizes reusable strategies and repair assets with validation records via the GEP protocol, offering market- and quality-based ranking to help different agents discover, evaluate, and inherit existing capabilities.

分类推荐榜首Pricing unverifiedView details
AI ShortWriting & content

AIS ALERT MANAGEMENT AND SHARED PLATFORM, QUICK COMMAND FOR MULTIPLE SCENARIOS

AI HINT COMMAND分类推荐榜首Pricing unverifiedView details
MemoAIMusic & audio

MemoAI is an AI-powered audio-to-text tool that supports localized processing, eliminating the need to upload data and ensuring privacy and security. It can efficiently transcribe YouTube videos, podcasts, and local audio/video files (such as MP4, MP3, AAC, M4A, etc.) into text, while offering core features like multilingual translation, speaker diarization, real-time subtitles, and floating notes—making it ideal for meeting minutes, interview transcription, content learning, and more.

AI AUDIO TOOL分类推荐榜首Pricing unverifiedView details
KiroCoding & development

Kiro is an AI engineering tool for developers that organizes requirements into specifications, architectures, and tasks, generates code via agents, and supports the entire workflow—from planning to code review—through property-based testing, execution memory, and integration with development tools.

AI programming tool分类推荐榜首Free entry availableView details
FinBusiness & support

CLIENT-ORIENTED AI SMART BODY。

Customer serviceSmartClient services待评分 · 分类列表推荐Pricing unverifiedView details
My Glow Up AILife & personal

My Glow Up AI provides free analysis based on selfies, without requiring registration, and can generate hairstyle recommendations and beauty plans that suit you.

待评分 · 分类列表推荐Pricing unverifiedView details
AssemblyAIMusic & audio

AssemblyAI is a speech AI platform for developers, offering pre-recorded, real-time, and short-form synchronous transcription APIs. It enables the building of voice-based applications by integrating capabilities such as speaker identification, emotion analysis, chapter and summary extraction, as well as speech agent functionality, dictation, and sensitive information processing.

AI AUDIO TOOL待评分 · 分类列表推荐Pricing unverifiedView details
ElicitKnowledge & research

Elicit provides semantic search, research report generation, literature screening, and data extraction for scientific research. It enables traceable answers within source materials and organizes evidence through a literature library, research agents, and ongoing tracking workflows, making it ideal for advancing from initial problem exploration to systematic reviews.

paperLiteratureResearch待评分 · 分类列表推荐Free entry availableView details

AI, as it happens

2409 published stories
Safety & governance · 154 storiesClear filters ×
2026-11-20
Follow the timeline
13:00

A Model of Diverse Sampling from Language Models

arXiv:2611.25003v1 Announce Type: new Abstract: Language models often produce high-quality individual samples but poor sets, with repeated samples clustering around the same semantic modes. We formalize diverse generation as sampling a size-$K$ subset of complete strings from a determinantal point process (DPP), where the likelihood encodes item quality, and the embedding geometry encodes repulsion between similar outputs. Exact inference over all strings is intractable, so we propose PoolDPP: draw a finite candidate pool from a tractable proposal, importance-weight the candidates toward the target quality, and run a $K$-DPP sampler on the induced pool kernel; a greedy MAP variant provides deterministic selection. We prove a finite-sample bound with an explicit constant: the KL divergence between the sampler's output and the target $K$-DPP decays as O(1/N) in the pool size. Empirically, PoolDPP extends the quality--diversity frontier on open-ended generation and the precision--recall frontier on ambiguous text-to-SQL (Ambrosia), and on mathematical reasoning (GSM8K), it raises diversity at about the same pass@$K$.

Copyright, privacy & ethicsRead original
2026-11-19
Follow the timeline
13:00

A Mechanistic Analysis of In-Context Learning and Out-of-Context Hallucination

arXiv:2611.24000v1 Announce Type: new Abstract: Large language models can infer and follow specific rules through in-context learning (ICL), yet may fail to remain faithful to provided evidence, giving rise to out-of-context (OOC) hallucination. Existing work has largely investigated the underlying mechanisms of ICL and OOC hallucination along separate lines. However, this separation obscures the mechanistic basis for these seemingly opposing outcomes. In this work, we study this contrast through the unified view of functional roles in dynamic context. The key observation is that role information can be internally represented without reliably constraining output. We derive a first-order decomposition to characterize the role effect, showing that its substantial influence depends on two necessary conditions, namely sufficient contribution magnitude and coherent signed aggregation. Causal analysis further shows that role span contributions affect key token choices involving response structure, content selection, and factual values. We further show that role-consistent choices depend on sufficient role span support at these key decisions, while failing choices can be reversed when this support overcomes the deficit. These findings support a common mechanistic view of ICL and OOC hallucination: each functional role defines the pattern of output decisions over which role span contributions must remain effective, and differences in this organization account for the seemingly opposing outcomes.

Copyright, privacy & ethicsRead original
13:00

Budgeted Proper Loss Design for Uncertain Decision Thresholds

arXiv:2611.23247v1 Announce Type: new Abstract: A binary probability estimate may be used by several decision makers with different error costs. Proper scoring rules can emphasize selected probability thresholds, but a single average threshold distribution need not give a good worst-case regret guarantee. We study the choice of a proper loss when the threshold distribution belongs to a finite specified family. Among integrable loss weights with a positive pointwise floor and a fixed total mass, we derive the weight that minimizes the sharp, distribution-free multiplier from surrogate regret to threshold-averaged decision regret. The solution clips a rescaled upper envelope of the threshold densities at the prescribed floor; its normalization is found by a one-dimensional root search. We give proofs of sharpness, optimality, and stability under density misspecification, as well as a limitation for atomic threshold laws. An exact histogram implementation and two synthetic experiments across 20 seeds verify the construction. The optimized multiplier improves over uniform and average-density weights, while fitted-model performance does not improve uniformly. The result is a loss-design guarantee under an explicit budget, rather than a general claim of predictive superiority.

Copyright, privacy & ethicsRead original
13:00

Coverage Governs Accuracy: When to Expand the Lexicon and When to Fall Back in Low-Resource Manchu OCR

arXiv:2611.23900v1 Announce Type: new Abstract: Character-level OCR for low-resource Manchu often produces near-miss spelling errors. Lexicon-constrained decoding can correct such errors with lexical knowledge, but a fixed dictionary cannot cover the many unseen word forms produced by agglutinative word formation. We study whether morphological expansion alleviates this limitation, and how the expansion size and confidence fallback affect recognition. We segment the training transcriptions with a dictionary-based stem list, extract roots and suffix chains from them, generate about 510k candidates, select the expansion size through frequency ranking on the validation set, and decompose accuracy under hard lexicon constraints into the product of coverage and in-lexicon precision. Across three CRNN training seeds, adding 250k candidates to a base dictionary of about 29k words raises the type accuracy on unseen combinations of seen roots from 63.6% to 86.3%, and overall token accuracy from 82.4% to 84.2%. Adding all candidates lowers overall token accuracy to 81.5%, showing that the coverage gain comes with a loss of in-lexicon precision. A further comparison shows that the base dictionary with confidence fallback performs best overall among the CRNN systems tested, reaching 76.1% type and 85.5% token accuracy and outperforming the expanded lexicon both with and without fallback. A Transformer recognizer shows similar trends. These results indicate that morphological expansion relieves the constraint of a fixed dictionary, but a larger lexicon does not guarantee better recognition; in the present low-resource Manchu OCR setting, a selective combination of reliable lexical constraints and open recognition is more effective.

Copyright, privacy & ethicsRead original
13:00

ExiST: Existence-Aware Self-Training for Semi-Supervised Multi-Organ Segmentation

arXiv:2611.23581v1 Announce Type: new Abstract: Semi-supervised multi-organ segmentation should learn more than a dozen organs from a handful of labeled scans, and the smallest organs occupy less than one percent of the foreground. We identify a failure mode of self-training that is specific to this regime. When a teacher misses a small organ in a scan, every voxel of that organ receives the confident pseudo-label background, and the student is trained toward the organ absence. Repeated over many scans, this loop extinguishes tail organs. In a pilot study, removing the gallbladder from the pseudo-labels reduced its Dice on AMOS from 36.5 to 1.0. We argue that a missing organ should be treated as unresolved rather than absent and propose Existence-aware Self-Training (ExiST). It builds pseudo-labels by leave-one-out consensus of independently trained networks, screens them with a quality gate on validation scans, and infers, per scan and per organ, whether the organ was found. For every unresolved organ, the background voxels of that scan receive the set-valued target, trained with a marginal cross-entropy that accepts either answer and a Dice term that skips uncertain voxels. We show that this loss never pushes a candidate organ toward absence and never favors one member of the set, which leaves the decision to labeled evidence. ExiST outperforms the current state-of-the-art methods and reaches 77.1 mean Dice on Synapse with 20% labeled scans and 75.9 on AMOS with 5% labeled scans.

Copyright, privacy & ethicsRead original
13:00

Grading Executable Legal Rules Against Independent Records: Transitional Provisions and Korean SME Status

arXiv:2611.24055v1 Announce Type: new Abstract: Laws are increasingly turned into programs that decide cases, such as whether a firm was a small or medium enterprise (SME) on a given date, and large language models (LLMs) now write them. The programs are usually graded against answers written by someone who read the same law, so a program can score well by sharing the grader's reading rather than by getting the law right. LexTemporis grades them instead against records of what actually happened: 10,901 firm-years of SME status disclosed by Korean listed companies, with every version of the governing rules. An appendix adds comparison cases from US income tax and German hospital minimum volumes. The date changes outcomes mainly through provisions outside the main statutory text: one transitional provision raises agreement with the disclosures from 83-86% to 89-90%. The disclosures also test the author's reading of a contested provision, favouring it in 130 of 189 firm-years where readings diverge. Four LLMs wrote the rules, given different selections of the law. A pre-registered comparison finds no condition better across all four models. In exploratory analyses, supplying the right provisions looks better graded against the author's reading than against the records: the advantage shrinks for three models (for one, to between a quarter and three-fifths of its size), and per firm-year it is several times larger where the author's reading is wrong, partly by construction, since there reproducing that reading more faithfully must lose. Grading legal programs against the grader's own reading can overstate how well they work.

Policy, regulation & standardsRead original
13:00

GUARD: Stabilizing Recursive Self-Improvement with Regret Guidance

arXiv:2611.23476v1 Announce Type: new Abstract: Autonomous AI agents pair a frozen base model with an editable harness of prompts, skills, tool rules, and memory. Recursive self-improvement (RSI) adapts this harness from experience without weight updates, yet self-editing without external graders often backfires: reflection and scalar scores conflate what an edit newly broke with what remains unsolved, causing local regressions, skill forgetting, and memory bloat. We introduce GUARD (Grader-free Unsupervised Environment Design [UED] with Adaptive Rule Distillation), a closed-loop framework for reference-free continual harness evolution. Our key insight is to decompose each task into pre-execution self-checks and guide harness updates via directional regret. Each round grades trajectories against these checks, pinning mastered rules while prioritizing newly broken behaviors, and selects among parallel candidate edits by critic-free group-relative fitness. A Soft-HURT gate then promotes an edit only when gains outweigh regressions, scoping specialized fixes to prevent over-triggering and guaranteeing non-regression on its gated buffer. At test time, one frozen harness runs in a single pass. Across benchmarks on memory, skill evolution, continual task streams, and domain reasoning, GUARD consistently improves performance without catastrophic forgetting or memory pollution while substantially cutting deployment token usage.

Copyright, privacy & ethicsRead original
13:00

Layer-Matched Sparse Attention for Low-Redundancy 3D Reconstruction

arXiv:2611.23554v1 Announce Type: new Abstract: VGGT reconstructs cameras and geometry from unposed images in one pass, but its global attention grows quadratically with the number of views. Training-free accelerators thin this attention and work well on redundant video, but lose pose accuracy when views share little geometry. This paper traces these losses to two kinds of cross-view information that a uniform sparse approximation can fail to preserve together: diffuse layers aggregate distributed context, whereas matching layers depend on query-specific interactions that carry little attention mass. A per-layer statistic measures the log-mass missed by a mean token. Its row-free proxy is the logit variance, and it selects the estimator for each layer. Diffuse layers use shared, importance-sampled hub keys, with log-weights that make the softmax numerator and denominator unbiased. Matching layers stay exact: every aggressive thinning tested cost accuracy. At equal cost, the selected layers outperform the four shifted windows tested. On 13 unseen low-redundancy scenes with 32 views, LMSA stays within 1.4 AUC@5 of dense attention, compared with 10.3 for AVGGT and 4.5 for SAF3R; its difference from contemporaneous GPA-VGGT is not significant, and only its gains over AVGGT pass both scene-level tests. With 100 views, it stays within 1.9 AUC@5 while its aggregator runs $2.0\times$ faster. Used post hoc to choose the layers GPA-VGGT protects, the statistic improves it by 1.2 AUC@5 for 5% more FLOPs. Applied untuned to $\pi^3$, it beats AVGGT but remains further from dense accuracy.

Copyright, privacy & ethicsRead original
13:00

RefinementGuard: Counterexample-Directed Semantic Qualification for Hidden Agent-Tool Execution

arXiv:2611.23611v1 Announce Type: new Abstract: Tool runtimes can lower an agent-selected operation into a hidden parallel graph, reducing latency and context while risking semantic mistranslation. We introduce RefinementGuard: translation validation for hidden, effectful agent-tool lowerings, coupled to artifact-bound, fail-closed runtime admission. Qualification compares a candidate graph with an independent reference on isolated state using explicit parent-operation obligations. We evaluate two repository-level contracts, bug localization and patch validation, on controlled repositories. Among 100 activated, non-equivalent development/qualification faults adjudicated by one human, parent-semantic validation rejects 49/100 (operator-cluster bootstrap 95% CI [33.7%, 64.8%]). The objective is selective admission, not maximum mutation rejection: exact equality rejects 100/100 faults but also all 60 valid variants, whereas obligation-based validation accepts 60/60; its 51 residual misses identify concrete specification gaps. On a frozen disjoint input split, the held-out B2-B3 result is localization-only because 21 patch mutants fail structurally; B3 rejects 38/56 localization defects. In a controlled asynchronous-delay regime, the localization serial/guarded median-latency ratio is 4.50x, and hidden execution exposes 74.3% less parent-visible context than visible execution. Patch validation shows no demonstrated speedup; zero-delay runs expose overhead. End-to-end qualification thus detects lower-layer misses while preserving localization's parallelism.

Copyright, privacy & ethicsRead original
13:00

SpecLook: Speculating Where to Look for Fast and Reliable Agentic MLLMs

arXiv:2611.23483v1 Announce Type: new Abstract: Agentic multimodal large language models (MLLMs) that think with images interleave reasoning with zoom-in tool calls. However, this sequential loop is slow, and its answers carry no reliability score. Prior speculation of tokens, reasoning steps, or answers keeps the loop, and answer speculation also accepts confident errors. We find that the loop mostly adds one zoom-in step, with which a small model is ahead of the agent on most splits. Our key insight is to speculate the agent's next observation, which, unlike a drafted answer, can be verified, corrected, and used to test the answer. We propose SpecLook, a training-free speculate-and-verify framework with three levels. A small model drafts a glimpse and re-answers from it, and the question ends if the answer is unchanged and confident. Otherwise, the agent's large model verifies and corrects the glimpse without tools, and their answers are fused. Only questions that still lack evidence reach the agent, in one forward pass that SpecLook-Fast skips, so no question runs the loop. The trace scores every answer's reliability without labels. Extensive experiments on V*, HR-Bench, and POPE show that both are more accurate than every baseline on every split. SpecLook reaches 87.2% and 86.2% mean accuracy with 7.7x and 4.7x speedups over the DeepEyes-7B and Thyme-RL agents, and SpecLook-Fast reaches 9.9x and 6.3x.

Copyright, privacy & ethicsRead original
00:56

Agents want the crown jewels: HPE and Daxa push governance down to the data layer

Enterprises are handing AI agents their most sensitive data, and governance is struggling to keep pace. Techniques such as semantic caching are emerging as a way to enforce controls inline without slowing agents down. Moving agents from pilot to production means letting them read, change and even delete the records that matter most. That shift […] The post Agents want the crown jewels: HPE and Daxa push governance down to the data layer appeared first on SiliconANGLE.

Copyright, privacy & ethicsRead original
2026-11-18
Follow the timeline
22:00

OneTrust’s Fall ’26 Release expands AI visibility, privacy and third-party risk tools

Governance platform provider OneTrust LLC today announced its Fall ’26 Release, a set of platform updates said to bring governance into the workflows where business decisions are made. Building on OneTrust CORIE, a shared intelligence layer for governing artificial intelligence agents introduced at the company’s TrustWeek conference in September, the release covers AI visibility, privacy, […] The post OneTrust’s Fall ’26 Release expands AI visibility, privacy and third-party risk tools appeared first on SiliconANGLE.

Copyright, privacy & ethicsRead original
13:00

DiscoveryAgent: A Framework for Scientist-centric Agentic AI

arXiv:2611.21950v1 Announce Type: new Abstract: AI agents can assist scientific research, but balancing autonomous investigation with scientists' ability to inspect and redirect ongoing work remains a central design challenge. We present DiscoveryAgent, a domain-agnostic, scientist-centric system that combines autonomous investigation with scientist-directed hypothesis steering. Scientists select, refine, or replace hypotheses between research rounds, while audit agents interrogate checkpointed records to trace claims, reproduce analyses, and inform subsequent decisions. Across two predictive modeling tasks, DiscoveryAgent produced the best variable-star classifier among evaluated baselines and, with scientist steering, a protein variant-effect predictor surpassed by only one of 106 ProteinGym baselines. We develop a rigorous checkpointing system that enables highly capable audit agents, a scalable form of oversight for agentic research. Audit agents autonomously identify errors at higher rates than previous AI scientist agents, and scientists collaborating with audit agents identify 14x more valid errors versus scientists performing manual review. In three case studies led by domain experts across physical science, social science, and life science, we developed a multimodal photometric-redshift estimator that resolves complementary failures of images and photometry, achieving 0.971 AUROC for low-redshift classification; documented an 83% rise in Florida homeowners-insurance premiums from 2013 to 2023, with a growing, regressive burden; and identified CD4+ T-cell relay genes, which both receive and exert broad transcriptional regulation, enriched for autoimmune disease associations. Together, these results show how autonomous research can expand scientists' empirical reach while centering their authority over research direction and their ability to understand and challenge the evidence.

Policy, regulation & standardsRead original
13:00

Latent Fact Composition under Next-Token Training

arXiv:2611.21921v1 Announce Type: new Abstract: Controlled studies of latent two-hop reasoning train transformers to predict only the answer, whereas language models are trained to predict every token. We ask whether a model trained on every token learns to answer two-hop questions whose intermediate entity appears in no two-hop training example. We build a synthetic corpus with such questions. On this corpus every term of the loss can be computed exactly. A looped transformer, a standard transformer and DiscoLoop answer almost none of these questions. We therefore introduce the Baton family of looped transformers. They pass the intermediate entity to the second loop as a token embedding. When trained to predict only the answer, two of its three variants answer these questions almost perfectly. We then relate the success and failure of each model to three design conditions. Experiments that change one condition at a time agree with this hypothesis. When Baton is trained on every token, however, its accuracy on these questions rises and then declines. We further split the gradient of the next-token loss into three terms. Removing the two terms that depend on the particular training set raises the peak accuracy but does not remove the decline. The decline arises in the second hop while the model still recalls both facts.

Copyright, privacy & ethicsRead original
13:00

On Formal Counterfactual Explanations

arXiv:2611.22010v1 Announce Type: new Abstract: Counterfactual explanations are often seen as counterfactual conditionals under closest-world semantics, requiring one to reason over all minimally distant counterexamples. In contrast, formal XAI typically relies on feature-based abductive explanations (AXps) and contrastive explanations (CXps), which admit efficient computation and a minimal hitting-set duality. This paper bridges these perspectives by (i) characterizing closest-world counterfactual explanations explicitly as a disjunctive normal form over all minimally distant counterexample instances, and (ii) introducing conjunctive counterfactual explanations (CFXps), a subclass of counterfactual antecedents restricted to conjunctions of feature literals that guarantee prediction change. The paper establishes structural results connecting CXps, AXps and CFXps: every CXp induces an AXp for its witness instance, and every CXp together with a witness AXp induces a CFXp whose support is precisely determined by that AXp. These theoretical results yield a modular approach for computing CFXps using existing AXp/CXp machinery. Experiments on a representative set of benchmarks using tree ensembles and neural networks show that CFXp extraction is practically efficient. Compared with popular point-based counterfactual explainers, CFXps generally require no more feature changes and can guarantee the counterfactual prediction beyond a single instance.

Copyright, privacy & ethicsRead original
13:00

Temporal Pseudo-RGB Encoding for Label-Free Detection of VEEV Infection in Human Brain Endothelial Cells

arXiv:2611.21892v1 Announce Type: new Abstract: Label-free detection of viral infection is central to antiviral screening and pathogen surveillance. Most deep-learning methods analyze individual frames and do not model temporal progression. We evaluate temporal pseudo-RGB encoding, which stacks three time-offset brightfield frames ($t{-}6$ h, $t{-}3$ h, $t$) as the red, green, and blue channels of a single image so that morphological change becomes color and an ImageNet-pretrained convolutional neural network (CNN) can be trained on infection trajectories without architectural modification. In human brain microvascular endothelial cells (HBMVEC) infected with Venezuelan equine encephalitis virus (VEEV) TC-83, a multi-task network predicts infected-versus-mock status and time post-infection in one forward pass. Across three experimental runs under leave-one-row-out cross-validation, with all 36 wells scored out of fold, the encoding reaches 91.6% accuracy against 87.4% for the single-frame model (paired difference $+4.2$ percentage points, 95% CI $[+2.5, +6.1]$; sign-flip $p = 3\times10^{-5}$; well-level inference, conditional on three runs). At the late-phase endpoint (40-48 h), where floating cells and debris can make conventional brightfield images difficult to interpret, the advantage widens to $+23.9$ points (95% CI $[+19.2, +28.9]$). Time post-infection is recovered on the fixed row split with MAE = 1.31 h ($R^2 = 0.975$). Among the temporal models tested the encoding has the highest mean accuracy, at a fraction of the cost of the recurrent and 3D convolutional alternatives, though only its margin over the single-frame model is resolved; the tested ViT-B/16 backbone showed no corresponding gain, so the finding is limited to convolutional backbones and this virus-cell system.

Copyright, privacy & ethicsRead original
13:00

TIBECG: Temporal Information Bottleneck for Lead-Specific Demographic Privacy in ECG Waveform Release

arXiv:2611.21694v1 Announce Type: new Abstract: Transmitting an electrocardiogram (ECG) waveform can expose demographic information beyond an intended diagnostic task. We evaluate TIBECG, an architectural temporal information bottleneck, on individual leads extracted from hospital twelve-lead recordings. A 2,302,337-parameter processor varies its latent length while retaining the same nine-score consistency objective, without age or sex labels in protector fitting. Across 6,386 recordings, ten folds, 12 leads and 21 lengths, retrospective selections retained at least 95% of each lead's clean macro-F1. Their weaker age/sex interval gain exceeded the best eligible conventional transformation on all leads, by a median 22.08 percentage points. The implemented sampling-rate comparison shows that direct downsampling usually suppresses demographic inference more strongly but at a larger loss of diagnostic utility, whereas TIBECG favors utility preservation. Selection stability also varies by lead. Lower latent Kullback-Leibler (KL) divergence does not reliably rank residual sex leakage at a fixed length. These findings support lead-specific evaluation of waveform processing, not validated consumer deployment or expert-readable reconstruction. When the waveform-input diagnostic classifier is hosted as part of a proprietary cloud service rather than deployed on the device, waveform transmission is required to obtain its scores.

Copyright, privacy & ethicsRead original
2026-11-17
Follow the timeline
13:00

Jailbreak LLMs with Linguistic Style as a Hidden Attack Surface

arXiv:2611.18601v1 Announce Type: new Abstract: We show that linguistic style is a systematic and previously overlooked attack surface for LLM safety. Across a diverse taxonomy of styles and multiple state-of-the-art models, stylistic rewrites of identical harmful goals produce order-of-magnitude differences in Attack Success Rate (ASR), with some styles achieving up to 80\% ASR while others remain below 6\%. To exploit this vulnerability, we introduce a lightweight single-pass attack framework that combines a style-conditioned generator with a BERT-based style selector, achieving performance competitive with substantially more expensive multi-turn optimization methods. To understand why style affects safety behavior, we conduct mechanistic analysis on LLaMA-3.1-8B and find that stylistic rewrites disrupt refusal representations while preserving recoverable harmful intent, suggesting a distributed failure mode rather than a single refusal mechanism. Motivated by this finding, we develop a style-aware Direct Preference Optimization defense that reduces ASR from 86\% to 19.5\% while largely preserving utility. Together, our results establish linguistic style as a critical factor in adversarial robustness and provide both mechanistic insight and a practical defense.

AI safety & riskRead original
Every update links back to its source.
STAY CURIOUS. STAY INFORMED.

Stay informed about the changes that matter.

Give your curiosity direction. Make every day of reading count.

Subscribe via RSS